Data Classification Algorithms Research Articles

Comparison of data augmentation and classification algorithms based on plastic spectroscopy.

Plastic waste management is one of the key issues in global environmental protection. Integrating spectroscopy acquisition devices with deep learning algorithms has emerged as an effective method for rapid plastic classification. However, the challenges in collecting plastic samples and spectroscopy data have resulted in a limited number of data samples and an incomplete comparison of relevant classification algorithms. To address this issue, we propose a plastic spectroscopy generation model and conduct a systematic analysis and comparison of different algorithms' performance from multiple perspectives, based on data augmentation. This paper first performs cubic interpolation, normalization, S-G filtering, linear detrending, and standard normal variate (SNV) transformations as preprocessing methods on plastic spectral data collected from public datasets using techniques such as Fourier Transform Infrared Spectroscopy (FTIR), Raman Spectroscopy (RAMAN), and Laser Induced Breakdown Spectroscopy (LIBS). The results, based on Principal Component Analysis (PCA) visualization, demonstrate that the preprocessing steps help improve classification accuracy. Additionally, PCA loading is used to explain the chemical classification features of each spectral device. Secondly, to tackle the issue of insufficient sample size, we propose a plastic spectroscopy generation model based on C-GAN, which effectively handles multi-class spectroscopy generation. The generated spectra are subjectively validated through difference spectroscopy and t-SNE to confirm their consistency with real spectra, and this conclusion is objectively validated using Maximum Mean Discrepancy (MMD). Finally, we compared the classification accuracy of machine learning algorithms, including Support Vector Machine (SVM), Back Propagation Neural Network (BP), K-Nearest Neighbors (KNN), Random Forest (RF), and Decision Tree (DT), with deep learning algorithms such as GoogleNet and ResNet under various conditions. The results indicate that after data augmentation using the plastic spectrum generation model, the accuracy of each classification model improved by at least 3% compared to pre-augmentation levels. Notably, for data collected via FTIR, the classification accuracy reached a peak of 0.991 under the 1D-ResNet model when the data were augmented twofold. For small sample datasets, traditional machine learning algorithms, such as SVM and RF, demonstrated high stability and accuracy, with only minimal differences compared to deep learning algorithms. However, on large sample datasets, deep learning algorithms showed a stronger advantage. Regarding data input formats, 1D input models generally outperformed 2D input models. Grad-CAM visualizations further illustrated that the 1D-ResNet model achieved the highest classification accuracy, primarily due to its ability to more accurately identify peak features in the data.

Sample Size Requirements for Popular Classification Algorithms in Tabular Clinical Data: Empirical Study

Background The performance of a classification algorithm eventually reaches a point of diminishing returns, where the additional sample added does not improve the results. Thus, there is a need to determine an optimal sample size that maximizes performance while accounting for computational burden or budgetary concerns. Objective This study aimed to determine optimal sample sizes and the relationships between sample size and dataset-level characteristics over a variety of binary classification algorithms. Methods A total of 16 large open-source datasets were collected, each containing a binary clinical outcome. Furthermore, 4 machine learning algorithms were assessed: XGBoost (XGB), random forest (RF), logistic regression (LR), and neural networks (NNs). For each dataset, the cross-validated area under the curve (AUC) was calculated at increasing sample sizes, and learning curves were fit. Sample sizes needed to reach the observed full–dataset AUC minus 2 points (0.02) were calculated from the fitted learning curves and compared across the datasets and algorithms. Dataset–level characteristics, minority class proportion, full–dataset AUC, number of features, type of features, and degree of nonlinearity were examined. Negative binomial regression models were used to quantify relationships between these characteristics and expected sample sizes within each algorithm. A total of 4 multivariable models were constructed, which selected the best-fitting combination of dataset–level characteristics. Results Among the 16 datasets (full-dataset sample sizes ranging from 70,000-1,000,000), median sample sizes were 9960 (XGB), 3404 (RF), 696 (LR), and 12,298 (NN) to reach AUC stability. For all 4 algorithms, more balanced classes (multiplier: 0.93-0.96 for a 1% increase in minority class proportion) were associated with decreased sample size. Other characteristics varied in importance across algorithms—in general, more features, weaker features, and more complex relationships between the predictors and the response increased expected sample sizes. In multivariable analysis, the top selected predictors were minority class proportion among all 4 algorithms assessed, full–dataset AUC (XGB, RF, and NN), and dataset nonlinearity (XGB, RF, and NN). For LR, the top predictors were minority class proportion, percentage of strong linear features, and number of features. Final multivariable sample size models had high goodness-of-fit, with dataset–level predictors explaining a majority (66.5%-84.5%) of the total deviance in the data among all 4 models. Conclusions The sample sizes needed to reach AUC stability among 4 popular classification algorithms vary by dataset and method and are associated with dataset–level characteristics that can be influenced or estimated before the start of a research study.

Algorithm for intelligent procesing of text information

The purpose of research. The goal of the research is to develop an intellectual processing algorithm for classifying text information. As the amount of information grows every day, it is necessary to quickly and efficiently separate significant from unimportant content. Therefore, the development of an intellectual processing algorithm for classifying text information is an urgent task.Methods. A method is proposed for classifying text information presented in one or more natural languages. It is based on 5 key stages: entering a task, accumulating a queue of tasks, processing the task, generating the result of processing the task, outputting the result. The input task is presented in the form of an http request, the body of which contains a file object. If the intensity of the input stream is greater than the processing speed, then an accumulation of tasks occurs. After selecting the active task (using the FIFO principle), it is processed. As a result of the transformations, the received data is decoded into a string using UTF-8 encoding. Processing refers to the process of categorization, when a search for patterns occurs in a line. Upon completion of rubrication, the result for the selected task is generated. From the accumulated result, a response to the original http request is formed, the body of which contains a list of found categories.Results. A method and algorithm for processing text data has been developed to determine the topics that are present in the input data set. The algorithm, implemented in software, allows you to work with text data in various languages.Conclusion. The software development of the text data classification algorithm was carried out in the C++ programming language using the Qt libraries version 5.11. This implementation showed a throughput of 1-5 MB per second (on a homogeneous input text data set). The algorithm allows you to correctly process damaged file formats.

Strategic Transformation of E-Commerce Big Data Classification and Mining Algorithms Based on Artificial Intelligence Era

Strategic Transformation of E-Commerce Big Data Classification and Mining Algorithms Based on Artificial Intelligence Era

Comparative Analysis of Fall Detection Using Artificial Neural Networks (ANN) and K-Nearest Neighbors (KNN) Algorithms

Fall detection is a critical area of research due to its implications for public health, particularly among older adults more vulnerable to fall-related injuries. This study explores the application of pervasive computing and machine learning algorithms in detecting falls using accelerometer and gyroscope sensors. Leveraging a dataset obtained from smartphone sensors, the study employs data processing techniques and classification algorithms, including K-Nearest Neighbors (KNN) and Artificial Neural Networks (ANN), to differentiate fall events from other activities. Evaluation metrics such as accuracy, precision, recall, and F1-score are utilized to assess the performance of the models. The results demonstrate the effectiveness of both the ANN and KNN models in accurately predicting fall detection with high precision and recall scores. The study highlights the potential of pervasive computing systems and machine learning methods to enhance fall detection capabilities, contributing to the development of proactive measures for ensuring the safety and well-being of individuals at risk of falls. Future research directions include exploring alternative algorithms and sensor modalities to improve fall detection systems further.

Improving the accuracy of honey bee forage class mapping using ensemble learning and multi-source satellite data in Google Earth Engine

In semi-arid agro-pastoral environments of Africa, beekeeping is widely recognized as an important activity to improve and diversify livelihoods. Although the scientific identification of suitable honey bees (Apis mellifera ssps.) forages may guide beekeepers to set up apiaries or to timely move honey bee colonies to exploit bee forage resources available in various landscapes, the characterization and mapping of bee forage classes is challenging. We evaluated how various data sources and classification algorithms in Google Earth Engine (GEE) affect the accuracy of honey bee forage class mapping in a semi-arid region of Ethiopia. Predictors derived from multi-source satellite data, such as high-resolution Planet imagery (P), Sentinel 1 RADAR (S1), Sentinel 2 multispectral (S2), and Shuttle Radar Topographic Mission (SRTM) Digital Elevation Model (DEM) were tested and best predictors were identified using Forward Feature Selection (FFS). Four machine learning algorithms (Gradient Tree Boost (GTB), Random Forest (RF), Classification and Regression Trees (CART), and Support Vector Machine (SVM)), all available in GEE, were compared and ensembled for honey bee forage class mapping. The results show that the highest accuracy is obtained by all four algorithms when combining P, S1, S2, and DEM compared to using predictors from a single data source or any other combinations. GTB had higher overall accuracy (90.9 %) than RF (88.2 %), CART (85.5 %), or SVM (79.9 %). Nonetheless, the highest overall accuracy (94.7 %) was obtained when integrating the four machine learning algorithms in an Ensemble Learning Approach (ELA). Applying ELA improved the classification accuracy by 3.8 %, 6.5 %, 9.2 %, and 14.8 % compared to single learner classification algorithms (i.e., GTB, RF, CART, and SVM, respectively). This study demonstrates an improved classification accuracy for honey bee forage class mapping in tropical rangeland by applying ELA, which can provide a better approach for monitoring and managing bee forage resources.

SnapE-ResNet: A novel electronic nose classification algorithm for gas data collected by open sampling systems

Increasing the depth of the neural network and implementing the ensemble of neural networks are two main methods to improve the accuracy of gas recognition algorithms. However, the issues of optimization difficulties or excessive resource consumption may occur when training deep neural networks or integrating multiple neural networks. In this study, an algorithm of snapshot ensemble is combined with Residual network (ResNet) to form a SnapE-ResNet which could simplify the training of the network due to the residual structure of ResNet, while realizes the ensemble of multiple ResNet models without additional time consumption and helps the network escape from local minimum using cyclic cosine annealing algorithm. The effectiveness of the proposed SnapE-ResNet is verified through classification experiments on the public dataset. Furthermore, the long-term drift and short-term drift are reduced by applying a zero-offset baseline compensation algorithm and a gaussian noise that added to the baseline signal. The experimental result indicating a high gas classification accuracy of 99.8 % from SnapE-ResNet is obtained, outperforming the comparative models. In addition, the necessity for snapshot ensemble is validated by network ablation experiments. This study could offer a reference for gas classification tasks in the gas detection fields.

Design of Time Series Fuzzy Segmentation Algorithm for Communication Data Classification

In order to solve the problem that the imbalance of communication data sets leads to a significant increase in classification difficulty, a classification algorithm for fuzzy segmentation of time series is proposed. The principal component analysis method is used to obtain the eigenvector with the largest eigenvalue. The time series of data is established according to the interval number theory. The segmentation target of the communication data time series is characterized by the Langmuir distance measurement function between data and categories. The classification result of fuzzy segmentation is obtained based on the judgment relationship between the difference of fuzzy classification matrix and the closing condition. The experimental results show that the number of data in the three classification situations of the algorithm in this paper is always at the corresponding ideal level, with high accuracy and low error and failure.

A novel binary data classification algorithm based on the modified reaction-diffusion predator-prey system with Holling-II function.

In this study, we propose a modified reaction-diffusion prey-predator model with a Holling-II function for binary data classification. In the model, we use u and v to represent the densities of prey and predators, respectively. We modify the original equation by substituting the term v with f-v to obtain a stable and clear nonlinear decision surface. By employing a finite difference method for numerical solution of the original model, we conduct various experiments in two-dimensional and three-dimensional spaces to validate the feasibility of the classifier. Additionally, with consideration for wide real applications, we perform classification experiments on electroencephalogram signals, demonstrating the effectiveness and robustness of the classifier in binary data classification.

On the gravitation‐based classification: A novel algorithm using equilibrium points for enhanced learning and dimensionality reduction

AbstractThe concept and effects of gravitation have been effectively utilized to design various data classification algorithms. Generally, there are two primary approaches to gravitation‐based classification: one that relies on gravitational force and another that is based on gravitational potential energy. In this paper, we examine these two approaches and introduce a novel classification algorithm grounded in gravitational potential energy. The core idea of our approach is to identify an equilibrium point for a line mass (serving as the classifier line) situated between two groups of fixed point masses (representing two data classes). The equilibrium point of the classifier line is determined by minimizing the total gravitational potential energy resulting from the two groups of point masses. Notably, our method demonstrates the following: (i) it acts as a dimensionality reduction technique that seeks a new feature space with lower dimensionality for improved class discrimination by maximizing the sum of the logarithms of the projections, (ii) it leads to an information‐theoretic learning strategy that minimizes the overall uncertainty of the classifier, and (iii) it offers a convex formulation that guarantees convergence to a global optimum solution. We also present experimental results that indicate the superior performance of the proposed method compared to existing techniques.

Low-cost and convenient screening of disease using analysis of physical measurements and recordings

In recent years, there has been substantial work in low-cost medical diagnostics based on the physical manifestations of disease. This is due to advancements in data analysis techniques and classification algorithms and the increased availability of computing power through smart devices. Smartphones and their ability to interface with simple sensors such as inertial measurement units (IMUs), microphones, piezoelectric sensors, etc., or with convenient attachments such as lenses have revolutionized the ability collect medically relevant data easily. Even if the data has relatively low resolution or signal to noise ratio, newer algorithms have made it possible to identify disease with this data. Many low-cost diagnostic tools have been created in medical fields spanning from neurology to dermatology to obstetrics. These tools are particularly useful in low-resource areas where access to expensive diagnostic equipment may not be possible. The ultimate goal would be the creation of a “diagnostic toolkit” consisting of a smartphone and a set of sensors and attachments that can be used to screen for a wide set of diseases in a community healthcare setting. However, there are a few concerns that still need to be overcome in low-cost diagnostics: lack of incentives to bring these devices to market, algorithmic bias, “black box” nature of the algorithms, and data storage/transfer concerns.

Expression of Concern: “Support vector machine and neural network for enhanced classification algorithm in ecological data”

Expression of Concern: “Support vector machine and neural network for enhanced classification algorithm in ecological data”

A hybrid Feature Selection Method Based on Binary PSO Algorithm for Microarray Data Classification

Microarray technology produces data with a large number of dimensions. The abundance of dimensions in the data makes it challenging for machine learning algorithms to extract meaningful information from it. To overcome this limitation, feature selection (FS) techniques can be applied to reduce the dimensionality of the data. FS is a crucial preprocessing step that allows for the handling of high-dimensional data and facilitates more effective and efficient processing. The primary objective of this paper is to develop an efficient FS method that can effectively reduce the dimensionality of microarray data. The Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) is utilized in the first filtering stage to extract informative features. The best feature subset is then determined by using a Particle Swarm Optimization (PSO) algorithm as a wrapper feature selector to identify ideal features. The proposed approach is evaluated using microarray datasets from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and compared with other algorithms. The experimental findings indicate that the proposed approach outperforms the benchmark methods in terms of classification accuracy.

Implementasi Model Support Vector Machine Dalam Analisa Sentimen Masyarakat Mengenai Kebijakan Penerapan Aplikasi Mypertamina

The policy for using the MyPertamina application issued does not rule out the possibility of differences of opinion due to changes in the policy. There are many positive, neutral, and negative responses to the MyPertamina application implementation policy. To see the public's reaction to the MyPertamina application implementation policy, it can be seen through various media, including social media. Twitter is a social network that is widely used by people in Indonesia. The number of Twitter users in Indonesia reached 18.45 million in 2022, making Indonesia the fifth largest Twitter user country in the world. Researchers conducted a sentiment analysis of the search results for tweets containing the keyword "MyPertamina" using the support vector machine algorithm. 382 tweet data were obtained and classified using the support vector machine algorithm. Support vector machine is a supervised learning algorithm for data classification. SVM is very fast and effective in solving text data problems. Text data is suitable for classification with the SVM algorithm because the basic nature of text tends to be high-dimensional. Of the 382 data analyzed, the support vector machine classification using the RBF kernel with parameter C=2 gave the highest accuracy value of 80.51%, precision value of 81%, recall value of 81%, and F1 score value of 80%.

Noncontact measurement of rectangular splines shaft based on line-structured light

Abstract The rectangular spline shaft, a typical type of shaft-tooth component, plays a significant role in mechanical transmissions. Existing methods for detecting size and positional tolerance in spline shafts often rely on contact-based measurement techniques, including specialized gauges and coordinate measuring machines (CMMs). To enhance the measurement efficiency, this paper proposes a method for measuring based on line-structured light. Firstly, a classification algorithm for data points on major and minor cylindrical surfaces and keyway surface of spline shaft is established, contributing to the automatic measurement. Secondly, a coaxiality error measurement model is established based on the overall least squares method, improving the measurement accuracy. Finally, a measurement model for key width and positional tolerance is established through the rotation of the spline axis. In experiments, the size and positional tolerance of the spline shaft obtained using this method are compared with measurements by CMM, meeting the general machining accuracy requirements.

Exploring the Impact of Artificial Intelligence and Machine Learning on Predicting Outcomes in Regional Development Conventions: A Case Study of the Tanger-Tetouan-Al Hoceïma Region, Morocco

Objectives: For over six years, Morocco has been engaged in an initiative aimed at promoting inclusive and sustainable territorial development, characterized by the implementation of an advanced regionalization experiment. However, the main challenge facing government leaders today is managing a growing number of regional development conventions. Methods: To achieve these objectives, we developed a predictive model using artificial intelligence and machine learning aimed at anticipating the outcomes of regional development conventions. Through this method, we were able to identify the risks associated with their potential failure and forecast their impact. Simultaneously, various data extraction methods and classification algorithms were implemented. We collected and analyzed information from 310 past and current regional development conventions, spanning from 2016 to 2020, considering various factors influencing their success or failure. We conducted extensive experiments to assess the effectiveness of various predictive models. Results: The results we obtained indicate that the classifiers Ada Boost, Random Forest, and Bernoulli NB are the three most effective algorithms for predicting the outcomes of regional development conventions. Conclusion: This study contributes to shaping a broader discourse on the practical application of artificial intelligence in public policies and regional development, highlighting its potential to enhance resource allocation and alleviate the burden of repetitive administrative tasks for organizations with limited resources.

Identification and Removal of Pollen Spectral Interference in the Classification of Hazardous Substances Based on Excitation Emission Matrix Fluorescence Spectroscopy.

Sensitively detecting hazardous and suspected bioaerosols is crucial for safeguarding public health. The potential impact of pollen on identifying bacterial species through fluorescence spectra should not be overlooked. Before the analysis, the spectrum underwent preprocessing steps, including normalization, multivariate scattering correction, and Savitzky-Golay smoothing. Additionally, the spectrum was transformed using difference, standard normal variable, and fast Fourier transform techniques. A random forest algorithm was employed for the classification and identification of 31 different types of samples. The fast Fourier transform improved the classification accuracy of the sample excitation-emission matrix fluorescence spectrum data by 9.2%, resulting in an accuracy of 89.24%. The harmful substances, including Staphylococcus aureus, ricin, beta-bungarotoxin, and Staphylococcal enterotoxin B, were clearly distinguished. The spectral data transformation and classification algorithm effectively eliminated the interference of pollen on other components. Furthermore, a classification and recognition model based on spectral feature transformation was established, demonstrating excellent application potential in detecting hazardous substances and protecting public health. This study provided a solid foundation for the application of rapid detection methods for harmful bioaerosols.

Pengimplementasian Tingkat Ketepatan Waktu Kelulusan Siswa (Studi Kasus Di MTS Nur Ibarhimy) Menggunakan Algoritma C4.5

Education has a very important role in shaping the individual and directing the development of society. As an educational institution, MTS Nur Ibrahimy has a responsibility to improve the quality and efficiency in the implementation of Education. MTS Nur Ibrahimy is located in Rantauprapat, Rantau Selatan district, Labuhanbatu Regency. MTs Nur Ibrahimy has been established since 2000 and has produced a number of students who successfully completed their education at this school. Along with technological advances, pattern exploration can be done by using data classification techniques obtained through the data mining process. Data mining is generally done because of the large amount of data, which can be used to generate patterns and useful knowledge in the business operations of a company. One of the methods developed in data mining is a way to dig up existing data to build a model, and then use the model to recognize other data patterns that are not contained in the stored database. In this context, a classification model is created to identify data patterns related to "Passed" or "not passed" status classes, based on pattern Determination results from training data. The Decision Trees Model is an implementation of the classification model in data mining. This Model builds a decision tree from training data consisting of records in a database. The C4.5 algorithm is one of the data classification algorithms that uses decision tree techniques and is able to manage numerical (continuous) and discrete data, and can handle missing attribute values. This algorithm produces rules that are easy to interpret. C4.5 has been tested in various classification cases, including in medical, trade, personnel, and various other fields.

Customer Satisfaction Analysis Using the C4.5 Algorithm at Askara Minimarket

Customer satisfaction is one way to improve the quality of service in a company, service satisfaction is also a factor in the company's development. If the company does not have quality service to the company, the company will not be able to develop and compete with other companies. The relationship between the company and consumers must run well, before employees enter the shop area, employees must be trained first. After the consumer has finished shopping, the consumer must choose whether he is satisfied or dissatisfied in the application, the results of the consumer's satisfaction and dissatisfaction will be entered into the data, this data will later be processed into Data Mining. The C45 algorithm is a data classification algorithm using a decision tree technique that is well known and preferred because it has advantages

A Statistical and AI Analysis of the Frequency Spectrum in the Measurement of the Center of Pressure Track in the Seated Position in Healthy Subjects and Subjects with Low Back Pain.

Measuring postural control in an upright standing position is the standard method. However, this diagnostic method has floor or ceiling effects and its implementation is only possible to a limited extent. Assessing postural control directly on the trunk in a sitting position and consideration of the results in the spectrum in conjunction with an AI-supported evaluation could represent an alternative diagnostic method quantifying neuromuscular control. In a prospective cross-sectional study, 188 subjects aged between 18 and 60 years were recruited and divided into two groups: "LowBackPain" vs. "Healthy". Subsequently, measurements of postural control in a seated position were carried out for 60 s using a modified balance board. A spectrum per trail was calculated using the measured CoP tracks in the range from 0.01 to 10 Hz. Various algorithms for data classification and prediction of these classes were tested for the parameter combination with the highest proven static influence on the parameter pain. The best results were found in a frequency spectrum of 0.001 Hz and greater than 1 Hz. After transforming the track from the time domain to the image domain for representation as power density, the influence of pain was highly significant (effect size 0.9). The link between pain and gender (p = 0.015) and pain and height (p = 0.012) also demonstrated significant results. The assessment of postural control in a seated position allows differentiation between "LowBackPain" and "Healthy" subjects. Using the AI algorithm of neural networks, the data set can be correctly differentiated into "LowBackPain" and "Healthy" with a probability of 81%.